1254
GENSLER AND B~un-o
VOL. 28
Synthesis of Unsaturated Fatty Acids. Positional Isomers of Linoleic Acid1 WALTERJ. GENSLERAND JOHN J. B R U N O ~ Chemistry Department of Boston University, Boston 15, Massachusetts Received November 1, 1966 Analogs of linoleic acid, an “essential fatty acid,” have been synthesized. Coupling 1-bromo-2-alkynes with the Grignard derivative of appropriate chloroalkylacetylenes in the presence of cuprous chloride formed three 17-carbon compounds, 1-chloro-7,lO-, 1-chloro-9,12-, and l-chloro-10,13-heptadecadiyne. Diisobutylaluminum hydride reacted with these diynes to give adducts, which on protonation yielded 1-chloro-cis,&-heptadecadienes. Adding another carbon atom by carbonating the Grignard reagents from the chloroheptadecadienes furnished the desired 18-carbon acids. In this way the following skipped (methylene-interrupted) unsaturated isomers of linoleic acid were obtained: cis-8,cis-ll-octadecadienoic, cis-lO,cis-13-octadecadienoic,and cis-ll,cis-14octadecadienoic acids. The acids were characterized by their melting points, indices of refraction, and the melting points of their tetrabromides. The 18-carbon compounds and their ozonolysis products were examined with the help of gas-liquid chromatography.
Only three of the approximately twenty octadecadienoic acids in the literature incorporate the methylene-interrupted or skipped unsaturated system. These acids, all biologically a c t i ~ e ,are ~ cis-9,cis-12-octadecadienoic, ci~-8,cis-ll-octadecadienoic,~ and cis-6,cis-goctadecadienoic acid.6 Only the familiar cis-9,cis-12 isomer (linoleic acid) has been synthesized.6 Since more information in this area is highly desirable, we undertook to synthesize and characterize three closely related isomers, the cis-8,cis-ll-octadecadienoic, cis-10, cis-13-octadecadienoic, and cis-ll,cis-14-octadecadienoic acids (IX). The present paper reports our results. Syntheses.-The same general approach served for all the syntheses. Each proceeded through a 17-carbon diacetylenic intermediate VII, which was half-reduced to the diethylenic analog VI11 and then extended to the final 18-carbon acid IX. The 17-carbon intermediates VI 1 were prepared by combining l-bromo-2alkynes (111) with chloroalkynes (VI). These starting materials were reached by brominating l-hydroxy-2alkynes (11) with phosphorus tribromide and by alkylating sodium acetylide with chloralkyl iodides (V). Combination of the fragments 111 and VI was effected by adding cuprous chloride as catalyst7,*to a mixture of the bromide I11 and the Grignard derivative of chloroalkyne VI. With ether as solvent, the yield of coupling product VI1 was as high as or even higher than with tetrahydrofuran. However, tetrahydrofuran mas preferred, since this solvent shortened the reaction time fifteenf old. We planned to arrive at the final dienoic acids by way of the corresponding 18-carbon diynoic acids, which should be available by carbonating the Grignard rea(1) Abstracted from the dissertation submitted by John J. Bruno t o the Graduate School of Boston University in partial fulfillment of the requirements for the degree of Doctor of Philosophy, 1960. The work was supported in part by research grant H 3773 from the National Heart Institute, U. S. Public Health Service. (2) Present address: Sprague Electric Co., North Adams, Mass. (3) Cf.H. J. Deuel, Jr., “The Lipids,” Val. 111. Interscience Publishers, Inc., New York, N. Y.,1957. (4) A. J. Fulco and J. F. Mead, J . Biol. Chem., 236, 3379 (1960). (5) A. J. Fulco and J. F. Mead, ibid., 284, 1411 (1959); W. Stoffel and E. H. .4hrens, Jr., J . L i p i d Res., 1, 139 (1960); J. F. Mead in K. Bloch’s “Lipide Metabolism,” John Wiley and Sons, Inc., Kew York. N. Y., 1960, p. 41. (6) The most recent synthesis is b y &de I.Gaudemaris and P. Arnaud, Bull. SDC. chim. Prance, 315 (1962). (7) Cf.W. J. Gensler and G. R. Thomas, J . Am. Chem. Soc., 73, 4601 (1951). ( 8 ) C/. W. J. Gensler and A. P. Mahadevan, ibid., 77, 3076 (1955). (9) See among others S. N. Ege, R. Wolovsky, and W. J. Gensler, abid., 83, 3080 (1961).
CHa(CHa)mC=CH I
C1(CHz),Cl IV
4
NaI
CH3(CHz),C=CCHZBr 111
HC=C( CHz),CI VI
‘\/ I
i
1.
2.
Mg
coz
gents derived from the 17-carbon diacetylenic chlorides VII. Actually, in one experiment, carbonation of the Grignard reagent from l-chloro-9,12-heptadecadiyiie (VII, m = 4,n. = 8 ) gave the desired 10,13-octadecadiynoic acid (40-.50%). But this experiment could not be repeated; thereafter all attempts to prepare the Grignard derivative failed. Although this remarkable inertness to magnesiumlO was not observed with l-iodo-9,12-heptadecadiyne,the main process with the iodo compound was Wurtz coupling. Since repla,cing the halogen in chloro compounds VI1 or in 1-iodo-9,12heptadecadiyne with a nitrile group was also unsatisfactory,“ this approach mas abandoned. We then turned to the 1111-VIII-IX sequence and so to the half-reduction of the triple bonds of chlorides (IO) Search of the literature revealed no examples of Grignard reagents prepared from chlorides containing acetylenic bonds. The unsuccessful attempts include trials with l-chloro-6,9-pentadecadiyne[W. R. Taylor and F. M. Strong, ibid., 72, 4263 (1950)], l-chloro-5-hexadecen-8-yne [L. Crombie and A. G. Jacklin, J . Chem. Soc., 1632 (1957)1, and I-chloro-4,7,10,13-nonadecatetrayne [.4. I. Rachlin, N. Wasyliw, and M. W . Goldberg, J . Ore. Chem., 26, 2688 (1961)l. Propargyl chlorides such as 1,4-dichloro2-butyne [A. W. Johnson, J . Chem. SOC.,1009 (1946)], 1-chloro-1,l-dialkyl2-alkyne [K. N. Campbell and L. T. Eby, J . Am. Chem. Soc., 62, 1798 (1940)], and I-chloro-2-heptyne (even with the entrainment method) [M. S. Newman and J. H. Wotis, ibid., 71, 1292 (1949)l also fail to react. (11) Sodium cyanide in alcohol gave tars, while cuprous cyanide in xylene [W. D. Celmer and I. A. Solomons, ibid.. 76, 3430 (195311 or in acetonitrile [B. C. L. Weedon, J . Chem. Soc., 4168 (1954)] gave only starting material.
MAY,
1963
POSITIOKAL ISOMERS OF LINOLEIC ACID
VII. Neither hydrogenation over a palladium-oncalcium carbonate ~ a t a l y s t nor l ~ ~reduction ~~ with a zinc-copper alloy in alcohol14 proved suitable. Diisobutylaluminum hydride, l5 on the other hand, reacted smoothly with the diynes to give the desired chlorodienes VIII. Whether necessary or not, we followed the practice of routiiiely reducing each diacetylene twice so as to eliminate the possibility of finding unchanged triple bonds in the dienes VIII. This method was used satisfactorily for aM the VI1 to VI11 conversions. The last step, conversion of the 17-carbon chlorodienes VI11 to the octadecadienoic acids IX, proceeded without difficulty. In sharp contrast to the reaction with chlorodiyne VII, a mixture of ethyl bromide16 and chlorodieiie VI11 reacted smoothly with magnesium. Carbonation of the Grignard reagent with gaseous carbon dioxide gave analytically pure acids IX in yields of 50-607, or better. The acids I X mere converted to stearic acid by catalytic hydrogenation so that any question of the presence of a branched chain could be dismissed. Each unsaturated acid had the double bonds in the assigned positions, since ozonolysis of the methyl esters led to the expected mono- and dicarboxylic acid cleavage fragments.” The infrared absorptions curves of acids IX and their methyl esters gave no indication of transethylenic materials ; estimates based on the absorption at 10.3 p limited the trans content to considerably less than 1%. The ultraviolet absorption at 233 mp showed that conjugated unsaturation was present a t a concenGas-liquid chromatographic tration lower than 17G.18 analyses of the methyl esters, by indicating homogeneities of 98-99910,19,20 were consistent with the results of the ultraviolet and infrared analyses.
Discussion With the exception of the 7,lO isomer, a series of skipped diunsaturated fatty acids from cis-6,cis-9through cis-ll,cis-14-octadecadienoicacid has now been reported. I n vivo fatty acid transformations can occur by dehydrogenating unsaturated fatty acids between the point of unsaturation and the carboxyl group and by extending the chain a t the carboxyl end.4,5,21Test of the generality of these processes by finding whether the new acids generate new families of skipped polyunsaturated acids is of considerable interest. The parent acids and their biosynthesized derivatives are also of importance in connection with essential fatty acid a ~ t i v i t y ,and ~ , ~even ~ more so in the controversial (12) H. Lindlar, Helu. Chim. Acta, 35, 446 (1952). (13) The degree of selectivity and stereospecificity in catalytic half-
hydrogenation of triple t o double bonds can vary. [Cf.N. A. Dobson, G. Elinton, M. Krishnamurti, R. A. Raphael, and R. G. Willis, Tetrahedron,
(1961).]
16, 16 (14) B.
S. Rabinowitz and F. S. Looney. J . A m . Chem. Soc., 76, 2652 (1953); A. J. Clarke and L. Crombie, Chem. Ind. (London), 143 (1957). (15) G. Wilke and H. Muller, Chem. Ber., 89, 444 (1956); also Ann., 629, 222 (1960). Note t h a t F. Bohlmann, E. Inhoffen, and J. Politt, ibid., 604, 207 (19571, using diethylaluminum hydride, failed to generate a n ethylenic bond from a n acetylenic bond, possibly because ether was used as solvent. 116) M. S. Kharasch and 0. Reinmuth, “Grignard Reactions of Nonmetallic Substances,” Prentice-Hall, Inc., New York, N. Y., 1954, pp. 3846.
(17) J. G. Keppler, Rec. trau. chim., 76, 49 (1957). (18) See R. T. Holman in “Methods of Biochemical Analysis,” Vol. I V , 1).Glick, ed., Interscience Publishers, Inc., New York, N. Y., 1957, p. 99. (19) Personal communications from R. A. Landowne. ( 2 0 ) R. A. Landowne and S. R. Lipsky, Biochim. Biophys. Acta, 4 6 , 1 (1961). (21) Cj. E. Klenk and H. Debuch, Ann. Rev. Biochem., as, 39 (1939).
1255
area of the relation between polyunsaturated acids, blood cholesterol, and atherosclerosis. 323 With several of the diunsaturated acids I X a t hand, comparison of some of their properties became possible. Table VI1 shows that the boiling points are closely bunched and that the indices of refraction are very similar. A clean separation of the acids by fractional distillation or a clear-cut distinction based on index of refraction appears unlikely. The melting points fall lower than those of the monounsaturated cis-octadecenoic The 8,11 acid has the lowest melting point, the 11,14 acid the highest. Our data are not precise enough to prove or disprove a generalization that the melting points go up as the unsaturation moves away from the carboxyl group or to show whether the zig-zag pattern observed for the melting points of the monounsaturated acidsz4 has its counterpart in the diunsaturated acids IX. Bromine added to acids IX to give crystalline tet,rabromides, whose melting points differ by no more than 7” (Table VIII). The tetrabromo derivatives froni adjacent members of the series show little depression in their mixture melting points. A tentative generalization may be drawn: only non-adjacent members of the series of tetrabromides will give appreciable depressions and ranges in their mixture melting points. Accordingly, caution is indicated before relying on the sharpness of melting point of a derived tetrabromo derivative as an indication of the absence of adjacent isomers. n‘one of the methyl esters of compounds I X showed more than one peak on gas-liquid chromatographic analysis. Table IX shows that the retention times on a 5ft. silicone column all lie between 28.2 and 29.4 min. (230’). h 5-ft. polyester packing permitted the esters to pass through more rapidly, retention t’imes of 1.55-19 min. (200’) being observed. Differences of 3-4 min. should permit a t least partial separation. Unexpectedly, the order of ret’ention time on the polyester column mas not regular, although the shortest time was observed for the 8,11 isomer and the longest time for the 11,14 isomer. Better separation was obtained with longer packed columns and especially with long capillary columns. 2O Homogeneity measures based on gas-liquid chromatography of the ozonolysis products indicated a 3-5.7g’i, content of positional isomers in the methyl esters of acids IX. These measures accordingly were not in line with the abovementioned gas-liquid chromatography results nor with estimates based on ultraviolet and infrared absorption. Actually, by the same kind of cleavage analysis, the methyl ester of Hormel “high purity” linoleic acid furnished dimethyl azeleat,e with no less than 9.7% in homogeneity. We feel that assays based on ozonolysis products are better regarded as minimum, instead of true, measures of homogeneity.25 ( 2 2 ) Cj. H . J. Thomasson. Intern. Reu. Vitamin Research, 26, 6 2 (1953). Also note “Essential F a t t y Acids,” Proc. Intern. Conf. Biochem. Problems of Lipids, Oxford, 1957 [Chem. Abstr., 6 3 , 17277 (1959)l. (23) Council Report, J . A m . Med. Assoc., 181, 411 (1962); J. Enselme, “Unsaturated F a t t y Acids in Atherosclerosis,” Pergamon Press, New York, N. Y., 1962. (24) W. F. Huber, J . A m . Chem. Soc., 7 3 , 2730 (1951). (25) Other indications t h a t the method of ozonolysis and analysis of the products is not free of complications have been found by Keppler,” R. R. .411en [J.O w . Chem., 21, 143 (1956)], F. L. Benton, A. A. Kiess, and H. J. Harwood [ J . A m . Oil Chemists’ Soc., 3 6 , 457 (1959)1, and E. Ucciani, J. Paaero, and M. Naudet [Bull.soc. chim. France, 1209 (1962)l.
1256
GEXSLERAND BRUXO
VOL. 28
TABLE I l-HYDROXY-2-ALKYNES (11) CHOs(CH2)mC=CCHzOH Yield, m
%
5"
72 82
-
Ca1cd.--
B . p . , OC.
n2'D
C
-Found----
H
C
H
62-63 ( 2 mm.) 1.4541 C~HWO f7.10 11.50 7;. 1 11.5 3o.c 80-81 (9 mm.) 1,4520 r2c I1 73-74 (15-16 mm.) 1.4507d CGHlaO i3.40 10.27 73.2 10.3 a Ch. Moureu and H. Desmots. Bull. soc. chim.. 131 27.360 (1902). W. J. Gender and G. R. Thomas, J . Am. Chern. Soc., 73, 4601 M. S. Sewman and J.'H. Wotia, ibid.,' 71; 1292 (1949). (1951). At 26".
TABLE I1 l-BROMO-2-ALKYNES (111) CHI( CH&,C=CCH*Br Yield, m
B p., QC.
n25D
%
67-68 ( 2 . 5 mm.) 1.4831 88 43-44 (4 mm.) 1.4878 82 ZC 76-78 (35 mm.) 1.4903 98 a Anal. Calcd. for CsH16Br: C, 53.20; H, 7.45; Br, 39.38. M. S. Newman and J. H. Found: C, 52.9; H, 7.3; Br, 39.7. M'otia, J . Am. Chem. Soc., 71, 1292 (1949), report b.p. 104-105' (56 mm.); W.J. Gender and G. R. Thomas, ibid., 73, 4601 (1951), report b.p. 76-80' (15 mm.) and TLD 1.491; J. H. Wotia, zbid., 72, 1639 (1950), reports b.p. 54" (4 mm.) and n% 1.4844. 31. S.Xewman and J . H. Wotiz, ibzd., 71, 1292 (1949), report b.p. 97-98' (80 mm.) and n Z 51.4884; ~ J . H. Wotiz, ibid., 72, ~ 1639 (1950), reports b.p. 38" (4 mm.) and n Z 51.4886. 5" 3b
Experimental26 I-Bromo-2-alkynes (HI).-Table I gives the results of combining appropriate acetylenic Grignard reagents with formaldehyde to form the intermediate propargyl alcohols (II).z7 Phosphorus tribromide with pyridine in catalytic quantities converted the propargyl alcohols (11) to 1-bromo-2-alkynes2*(See Table I1). One-mole quantities could be handled without difficulty. Chloroalkynes (VI).-The chloroiodoalkanes (V) were prepared from the corresponding d i c h l o r ~ a l k a n e s . ~Table ~ ~ ~ I11 summarizes the data. The directions used for alkylating sodium acetylide with the chloroiodoalkanes to give chloroalkynes VI are an adaption and modification of a procedure reported before for 9-chl0ro-l-nonyne.~~The results in Table IV were obtained when sodium acetylide from 0.66 mole of sodamide was allowed to react with 0.26 mole of chloroiodoalkane in 100 ml. of ether plus 1.2 1. of liquid ammonia. 1-Chloroheptadecadiynes (VII).-The three diynes (VII) were obtained from the appropriate 1-bromo-2-alkyne (111)and chloroalkyne ( V I ) by following essentially the same procedure, which is given below for 1-chloro-7,lO-heptadecadiyne (VII, m = 5 , n = 6). The reaction was carried out under dry nitrogen in a carefully dried three-necked flask equipped with dropping funnel and magnetic stirring bar. The tetrahydrofuran solvent was purified by treatment with potassium hydroxide and then distillation from lithium aluminum hydride. A solution of ethyl Grignard reagent was formed by dropping 20 ml. (23.0 g.; 0.211 mole) of dry ethyl bromide in 50 ml. of tetrahydrofuran into a stirred mixture of dry magnesium shavings (5.07 g.; 0.211 g.-atom; Dow sublimed metal) and 150 ml. of tetrahydrofuran. Bringing the mixture initially to a boil started the reaction, which then proceeded without external heating during the remainder of the addition. hfter another hour of boiling, the Grignard solution was treated with a solution of 29.25 g. (0.201 mole) of 8-chloro-loctyne (VI, n = 6 ) in 50 ml. of tetrahydrofuran. The dropwise addition required 0.5 hr. The spontaneously boiling mixture (26) Analyses were performed by Carol K. Fitz, Needham Heights, Mass. (27) Cf.Taylor and Strong, J . A m . Chem. Soc.. l a , 4263 (1950); Tchao Yin Lai, Bull. soc. chzm., I41 63, 682 (1933); R. A. Raphael and F. Sondheimer, J . Chem. SOC.,2100 (1950). (28) hl. S. Newxnan and J. H. Wotiz, J . A m . Chem. Soc., 71, 1292 (1949). C / . Taylor and Strong2' as well as Tchao Yin Lai, ref. 27 and Bull. soc. chim., [ 4 ] 63, 1533 (1933). (29) C / . R. A. Raphael and F. Sondheinier.2' (30) W. J. Gender and C. B. Abrahams, J . A m . CAem. Soc., 80, 4593 (1958).
evolved a gas. The reaction mixture was boiled for an additional hour and then allowed to stand overnight under nitrogen. Dry cuprous chloride (0.7 9.) was added, and the mixture wag boiled for 1 hr. 1-Bromo-2-nonyne (111, m = 5 ; 40.8 g.; 0.201 mole) in 50 ml. of tetrahydrofuran was added to the stirred solution, which was then boiled for 2-4 hr. Approximately 887, of the initial Grignard content was consumed in the first hour, 94% after 2 hr. The reaction mixture, containing a heavy, bright green precipitate, was poured over 500 g. of crushed ice plus 50 ml. of concentrated sulfuric acid. The aqueous layer was extracted with several 100-ml. portions of ether. The combined organic phases were washed with water until the washings were neutral to litmus and then dried with magnesium sulfate. Removal of solvent left the oily product, which was fractionated through a 5-cm. Vigreux column. The water-white l-chloro-7,10-heptadecadiyne (VII) was distributed in ampoules, which were sealed without releasing the vacuum. In some preparations the distilled material contained small amounts of terminal acetylenic impurities as shown by a small absorption peak a t 3.04 p . Such impurities were readily precipitated by shaking the diyne (10 9.) with a saturated methanolic solution (50 ml.) of silver nitrate. Table J' presents the data for the I-chloroheptadecadiynes. The infrared absorption spectra of the three compounds, taken as neat layers, %-erepractically identical. The major absorption peaks occur a t 4.35 (vw), 4.42 (vw), 4.48 (vw), 6.8 (s), 7.61 (s), 13.8, and 15.35 p (9). S o peaks were noted a t 3 p (acetylenic hydrogen) or a t 5.1 p (allenic unsaturation). 1-Chloroheptadecadienes (VIII) by Diisobutylaluminum Hydride Reduction of 1-Chloroheptadecadiynes (VII).-The reagent was prepared31 by boiling 160 ml. of a 257, solution of triisobutylaluminum in heptane for 2-3 hr. Distillation afforded 27 ml. of colorless diisobutylaluminum hydride, b.p. 80-90' (0.05 mm. j The recommended precautions for handling organoaluminum compounds were ob~erved.15332 The reduction of l-chloro-7,10-heptadecadiyneis given here as illustrative. The diyne (13.76 g.; 0.052 mole) was placed in a 300-ml. three-necked flask fitted with a mercury-sealed stirrer, a vertical condenser, and a dropping funnel. Dry, oxygen-free nitrogen, introduced through the top of the condenser under a slight positive pressure, blanketted the reagent and the reaction mixture a t all times. Approximately 21 g. (0.15 mole) of diisobutylaluminum hydride was added over a period of 1 hr. to the stirred mixture a t 0 " . After residual reagent was rinsed into the flask with a small volume of dry heptane, stirring was continued a t room temperature for 12-15 hr. The condenser was replaced with a tube leading to a trap held a t Dry Ice-acetone temperature. With the stirred reaction mixture in a bath a t - 5 " , 60 ml. of a methanol-petroleum ether (b.p. 30-60') solution (2:3) was carefully added in 1 hr. Then just enough ice-cold 207, sulfuric acid was slowly added to dissolve the precipitated aluminum methoxide. After the condensate in the trap was transferred to the acidified mixture with 50 ml. of ether, the aqueous layer was separated and extracted twice with ether. The combined organic layers were washed in succession with water, saturated sodium bicarbonate solution, and again with water until the washings were neutral to litmus. The solution was dried with magnesium sulfate and then warmed on the steam bath under water-pump vacuum to remove all solvent. To eliminate all possibility of acetylenic impurities, the oil was treated routinely with a second portion of reagent in a repetition of the above procedure. (31) K. Ziegler, Chem. Abatr., 61, 15081 (1957) [British Patent 778,098 (1957)l; K. Ziegler, H. G. Geller, H. Lehmkuhl, W. Pfohl, and K. Zosel, A n n . , 629, l ( 1 9 6 0 ) . (32) C / . "Handling and Properties of Triisobutylaluminuin," Hercules Powder Co., Wilmington, Del.
1257
PORITIOSAL ISOMERS OF LIXOLEIC ACID
MAY,1963
TABLE I11 CHLOROIODOALKANES (V) I(CHz),Cl -Calcd. B.p., "C.
n
6" 8b
Foilnd----
H
C
n%
C
I
H
I
55.5-57 (0.25 mm.)
1.5176 85-89 (0.25 mm.) 1.5089 CsHisClI 35.00 5.88 46.20 35.0 5.8 46.4 nc 96-98 (0.25 mm.) 1.5066 CQHisClI 37.50 6.28 43.95 37.7 6.3 43.8 a W.J . Gender and G. R. Thomas, J . Am. Chem. SOC., 73,4601 (1951), report b.p. 112-116' (12 mm.) and nZ5D1.5220; R. A. Raphael and F. Sondheimer, J . Chem. Soc., 2100 (1950), report b.p. 73-74' (0.7 mm.) and nZ4D1.5248; W. F. Huber, J . Am. Chem. Soc., 73, 2730 (1951), reports b.p. 96-98' (6 mm.) and n Z 51.5214. ~ W. F. Huber, ibid., 73, 2730 (1951), b.p. 101-105' (2.5 mm.), and nZ51) 1.5113. K. Ahmad, F. M.Bumpus, and F. M. Strong, ibid., 70, 3391 (1948), report b.p. 123-124' (2.8-2.9 mm.) and n Z 5 D 1.5060; W.F. Huber, ibid., 73, 2730 (1951), reports b.p. 123-126' (4 mm.) and 7tZ5D 1.5074.
TABLE IV CHLOROALKYNES (VI) HC=C( CHz),C1 Yield, n
%
Calcd. B . p . , OC.
C
n2bD
H
c1
.
F-ou-n-d----
C
H
c1
6" 77 4&42 ( 2 . 5 mm.) 1.4507 CaH13Cl 66.40 9.06 24.50 66.3 9.1 24.3 8 88 56-59 ( 2 . 5 mm.) 1.4528 CioHiTCl 69.60 9.93 20.55 69.6 9.8 21.4 9 80 57-58 (0.15-0.2 mm.) 1.4538 CllHld21 70.80 10.25 70.9 10.5 a R. A. Raphael and F. Sondheimer, J . Chem. Soc., 2100 (1950),report b.p. 73-76" (10 mm.) and n i 21.4590; ~ W. J. Gender and G. R. Thomas, J . Am. Chem. SOC.,73, 4601 (1951), report b.p. 73-76' (11 mm.) and n Z 5 D 1.4548.
TABLE V 1-CHLOROHEPTADECADIYNES (VII) CHI( CHz)rnC=CCHzC=C( CHz)nCI Yield, -Founda--m 5 3 2
n tj
8 9
B.p., OC. 103-107 ( 0 . 5 X 10-1mm.) 106-109 (1 X 10-:mm.) 101-111 (0 5 x 10-:mm.)
%
n% 1.4784 1.4779 1.4780
70 68* 62
C1 C H 7 6 . 6 10.2 1 3 . 5 7 6 . 6 1 0 . 3 13 1 76.7 10.3 13.2
'
Calcd. for C17Hz7C1:C, 76.51; H, 10.20; C1, 13.29. With a 96-hr. reaction period in ether solvent, the yield was 81%. The light yellow product (13.0 g.) dissolved in petroleum ether (b.p. 30-60") was placed on a column of 130 g. of 28-200-mesh silica gel (Davison Chemical Co.). The first 250 ml. of eluate (petroleum ether solvent) contained 0.4 g. of a fatty solid which was discarded. The next 500 ml. of eluate (petroleum etherether solvent, 4: 1) furnished 11.4 g. (82%) of pale yellow, solventfree product. Distillation afforded colorless t o faintly yellow l-chloro-7,10-heptadecadiene(VIII), which was collected and stored in UQCUO in sealed vials. The 9,12 and 10,13 isomers were prepared in a similar way as colorless oils. The three isomers all showed very similar infrared absorption curves, with peaks evident a t 3.32 (m), 6.05 (w), 7.15 (m), 7.26 (m), 7.64 (m), 7.8 ( m ) , 10.3 (vw), 11.0 (m), 13.8 (s), and 15.3 (s) p . The absorption curve of the 7,lO isomer was taken before as well as after distillation. The close similarity in the two curves suggested that the chromatographed material was practically pure, and that distillation (presumably with the other isomers as well), with its accompanying losses, may have been unnecessary. Table VI presents the data.
TABLE VI ~~~,&CHLOR~HEPTADECADIENES (VIII) CH,( CH*),CH=CHCHzCH=CH(
CHz)nCl
Yield, 7%-
m n 5 6 3 8 2 9
B.p.. O C . 87-91 (1 X 10-4 mm.) 87-89 (1 X lO-4mm.) 75-77 ( 0 . 5 X lO-4mm.)
n% 1.4683 1.4678 1.4678
a
b
65 83 70
82 94 87
After distillation. Before distillation. H31C1: C, 75.37; H, 11.54; C1, 13.09.
--FoundcC H 75.8 75.3
C1
11.8 12.7 11.5
Calcd. for Cn-
&,cis-Octadecadienoic Acids (IX) .-The appropriate l-chloroheptadecadiene was converted to its Grignard reagent and then carbonated. The procedure, the same for the three acids, is given below for 8,ll-octadecadienoic acid (IX, m = 5 , n = 6).
A 250-ml. three-necked flask was fitted with a vertical condenser. ??, oxygen-free nitrogen, which was supplied under a slight positive pressure through the top of the condenser, blanketted the reaction during the entire experiment. All glassware was carefully dried. Turnings of Dow sublimed magnesium (2.28 g. or 0.094 g.-atom) were dried briefly in the flask by brushing the flask with a soft flame. A mixture of I-chloro-7,lO-heptadecadiene (VIII, m = 5, n = 6) (5.00 g.; 0.0184 mole) and carefully dried ethyl bromide (6.05 g.; 0.0555 mole) was placed in a dropping funnel and was diluted with 45 ml. of ether distilled directly into the dropping funnel from lithium aluminum hydride. Approximately 80 ml. of ether, distilled directly into the flask, covered the magnesium. The ethereal halide solution was added to the boiling, magnetically stirred mixture over a period of 2 hr. After the addition, the mixture was boiled for 18 hr. Gaseous carbon dioxide, after pawing through a tower of concentrated sulfuric acid and then through an empty flask, was introduced into the cooled mixture through a tube reaching almost to the bottom of the flask. The inside temperature, initially -60°, climbed t o - 5 " and then rapidly dropped t o -60". Solidification was noted. Enough cold 10% sulfuric acid was added to dissolve the excess magnesium. The aqueous layer was separated and extracted with several portions of ether. The combined ether layers were shaken with several portions of water and then dried with magnesium sulfate. All material volatile below 50-60" (ca. 0.5 mm.) was removed, and the yellow residual oil (5.07 g.), diluted w-ith an equal volume of light petroleum ether, was placed on a 2.5 x 33 cm. column of 28-200-mesh silica gel (100 g.). Passing 300 ml. of petroleum ether through the column removed 0.3 g. of material, which was discarded. The solvent was changed to petroleum ether-ether (4: 1, v./v.), and 500 ml. was collected. Removal of solvent left 4.23 g. of 8,ll-octadecadienoic acid, which was distilled through a 5-cm. vacuum-jacketed Vigreux column. The practically colorless main fraction was distributed in several small ampoules, which were sealed without breaking the vacuum. Melting points were determined by observing 0.1-0.2 g. of sealed solidified material as it was allowed to warm slowly in a bath. The infrared absorption curves of the acids IX, taken with neat samples, were all very similar and corresponded closely to the curve for "high purity" linoleic acid (Hormel). All showed absorption peaks a t 3.76 (m), 5.85 (s), 6.04 (w, sh), 7.17 (m), 7.26 (m), 7.8 (s), 8.1 (m), 10.7 (s), and 13.85 (s) p . Ultraviolet absorption curves were determined with ca. 1.8 X A{ methanolic solutions of the acids. The extinction a t the 233 mp maximum (log e 2.2-2.5) served as a measure of the amount of conjugated impurities.l* Neutralization equivalents were determined by titration of the acids in 957, alcohol under nitrogen with 0.05 N aqueous sodium hydroxide to a phenolphthalein end point. The values were slightly high (0.43-0.9070); but pure stearic acid by the same procedure also gave high results (0.18-
GENSLERASD BRUNO
1258
VOL. 28
TABLE VI1 C ~ S ~ C ~ S - ~ C T A D E C A D I E N OACID IC (IX)
CH3(CH2)mCH=CHCHzCH=CH( CHz),COOH m
Isomer
B.p., OC. (10-4 mm.)
n
6 7
Yield,a iMp., O C . (approx.)
A9,12
-12.5 to - 9 . 5 - 5 . 1 to -5.4e
1.4663 1.4690e
Conjugation, %
58
0.62 0.45
7FoundbC
77.1
H
Neut.c equiv.
11.5
283
Iodined no.
185,186 185,185 184’ A10,13 3 8 121-128 -9to -6.5 1.4670 51 0.87 77.2 11.5 282 184 9 112-115 6 . 0 to 8 . 0 1.4664 60 0.73 77.3 11.5 282 185 2 a The figures refer to distilled material. The chromatographed arids were obtained in 20-30% higher yield. Much of the decreased vield of distilled material is due to mechanical losses. Calcd. for C&202: C, 77.09; H, 11.50. e Molecular weight: 280.44. Theoretical iodine number: 181. e Reported by B. Sreenivasan, J. B. Brown, E. P. Jones, V. L. Davidson, and J. Nowakowska, J . Am. Oil Chemists‘ Soc., 39, 255 (1962). The index of refraction refers to 20’. Synthetic linoleic acid was reported recently with b.p. 115~ [M. de Gaudemaris and P. Arnaud, Bull. soc. chim. France, 315 (1962)l. ’The Wijs 120” (0.001 mm.), m.p. - S o , and n Z 21.4670 iodine number of this Hormel “high purity” linoleic acid was found by the supplier to be 181, the theoretical value. 5 4
A8.11
118-120
%
?? 2bD
’
0.7%). Iodine numbers were obtained by the Kaufmann method.33 Table VI1 summarizes the results. Hydrogenation of cis,&-Octadecadienoic Acids (IX) to Stearic Acid.-A mixture of the unsaturated acid (ca. 0.1 g.), platinum oxide (ca. 0.06 g.), and 95% alcohol (20 ml.) was magnetically stirred a t room temperature in an atmosphere of hydrogen until absorption of hydrogen ceased. After filtration, the solution was concentrated under reduced pressures to a volume of 2-3 ml. Cooling the concentrated solution gave white crystals, which were collected and dried in vacuo. Stearic acid, melting within the range 68-70’, was obtained in better than 9776 yield from each of the three acids IS. Mixtures of the hydrogenation stearic acids with authentic stearic acid (m.p. 69.5-70”) melted a t 68-70’. Tetrabromostearic Acids from Skipped Octadecadienoic Acids (IX).-Bromine was added dropwise to a stirred solution of 0.10.5 g. of unsaturated acid in 6-7 ml. of low boiling petroleum ether a t temperatures held below 0”. When the yellow color persisted, the precipitated solids were collected and washed on the funnel with cold petroleum ether. The compounds were recrystallized two or three times from methylene chloride-pentane or methylene chloride-cyclohexane and then dried in vacuo a t 56”. Table VI11 summarizes the results. From the reported melting point of linoleic acid tetrabromide (m.p. 113.2-113.8°34; 115.4-115.5°5), we judge that the melting points in Table VI11 might be slightly low.
TABLE VI11 DERIVATIVES OF OCTADECADIENOIC ACIDS (Ix) CH3(CHz),CH-CHCH,CH-CH( CHz)nCOOH
mm. thicknesses of the neat oils. All showed maxima a t 3.33 (m), 3.49 (m), 6.05 (w), 7.15 (m), 7.26 (m), 7.34 (m), 7.58 ( w ) , 8.05 (m), 8.36 (s), 8.56 (s), 9.85 (m), 10.1 (w), 10.98 (w), and 13.85 (m) p . The minor differences in the 8-9- and 11-12.5-p region offer little encouragement in the use of infrared to distinguish the several esters. Except for these differences, all the esters including methyl linoleate showed practically the same curves. The curves showed no characteristic trans double bond absorption a t 10.3 p . Since a distinct shoulder was observed here when a test mixture of methyl linoleate containing lye of methyl trans-11octadecenoate was scanned under the same conditions, the content of trans double bonds in the synthetic compounds was substantially less than 1%. Gas-Liquid Chromatographic Analysis of the Octadecadienoic Acids I X and Esters.-Table I X gives the results of gas-liquid chromatography with the methyl octadecadienoates. Only one peak appeared for each ester. The retention times for the esters on a column with a polyester stationary phase show significant differences, but the trend of retention time with the position of the skipped unsaturation proved not to be regular.
TABLE IX GAS-LIQVIDCHROMATOGRAPHIC ANALYSIS@ OF THE METHYL ESTERS OF OCTADECADIENOIC ACIDS(IX) CH3(CHz)mCII=CHCH2CH=CH( CHZ),COOCH3
TETR.4BROMO
I
Br Starting isomer
n
m
/
Br
I
1
Br
Br
Yield,
M.p.,a
-Foundb-
%
OC.
C
H
Br
5 6 20 106-107 35.9 5 . 2 4 i 35 111.5-112.5 36.3 5 . 1 53.3 Alo,’3 3 8 22 111-111.5 36.1 5 . 1 53.4 112-112.5 36.2 5 . 2 53.5 9 26 2 9,121 a The mixture melting points are as follows: [8,11 106-107”; [9,12 10,131 107.5-112.5’; [9,12 11,141 9498.5’; [10,13 11,141 108-11lo. Calcd. for C18H32Br402: C, 36.02; H , 5.38; Br, 53.26. Hormel “high purity” linoleic acid.
+
m
n
column-Polyester Retention Temp., “C. time, min.
columnRetention time, min.
5 6 231 29.4 201 15.5 As.” 4 7 230 29.0 201 18.3 28.2 201 15.6 8 231 3 19.1 2 9 235 28.6 201 a An “Aerograph” unit with catherometer detector was used with two 5-ft. columns, one packed with silicone(samp1e size, 10 pl.) and the second with a polyester (“LAC-446”; sample size, 1-3 pl.). The carrier gas was helium a t a constant flow Methyl ester of Hormel rate (rotameter reading, 85.0 mm.). “high purity” linoleic acid. A’,’’
As,’’ Ae,lzc
+
Isomer
--Silicone Temp., OC.
+
+
Methyl cis,&-0ctadecadienoates.-The acids I X were esterified by treating 0.1-0.2 g. samples in 10-15 ml. of dry ether with distilled ethereal diazomethane until a yellow color persisted. An atmosphere of pure nitrogen was maintained in the flask. After filtration through thin layers of a filter aid (Celite), anhydrous magnesium sulfate, and activated charcoal, the solution was warmed in a stream of nitrogen to remove solvent. The residual colorless esters in several vials were exposed without delay to a 0.01-mm. vacuum. The vials, sealed without breaking the vacuum, were stored in the cold away from light. Infrared absorption curves of the esters were taken with 0.04(33) N. D. Cheronis and J. B. Entrikin, “Semimicro Qualitative Organic Analysis,” 2nd ed., Interscience Publishers, Inc., New York, N. Y..1957, p. 736. (34) H. M. Walborsky. R. H. Davis, and D. R. Howton, J . Am. Chem. Soc., 75, 2590 (1951).
’
Longer packed columns and, even better, capillary columns gave more effective separation. These results are fully documented e1sewhere.M With the more efficient columns, the four methyl esters showed 98-9970 homogeneity. The maximum single impurity was observed in the methyl l i n ~ l e a t e . ~ ~ v * ~ After approximately 6 months of storage a t -so in vacuo away from light, the acids themselves were examined. The analyses were performed a t the Unilever Research Laboratory (T’laardingen) through the courtesy of Dr. R. K. Beerthuis. A column (120 cm. long; 0.4-em. diameter) packed with 10% Apiezon L on diatomaceous earth (Celite) was used. At a flow rate of 20 ml. per min. of argon and a t a column temperature of 182” the retention times of the 18-carbon acids were just over 2 hr. The column gave only partial separation of the 8,11 isomer (with the shortest retention time) from the other acids. Homogeneity estimates were made as follows: for the 8,11-acid, ca. 945%; for the 10,13 acid, ea. 100%; for the 11,14 acid, ca. 96%. Whether the appearance of inhomogeneity in the 8,11 and 11,14 isomer is
M . ~ Y1963 ,
IDEXTIFICATIOX OF MICROMEROL .4s URSOLIC ACID
1259
TABLE X GAS-LIQUIDCHROMATOGRAPHIC ANALYSIS~ OF
THE
OZONOLYSIS METHYL ESTERSDERIVED FROM OCTADECADIENOIC ACIDS -Retention
Octadecadienoic acid IX Isomer m
5
Column temperature, 100’
n
6 7 8 9
Reference methyl ester
time, min. Column temperature, 190‘
Cleavage ester
Reference methyl ester
Cleavage ester
Impurity,
%
7.25 2.9 Octanedioate 7.50 A9,1Zb 4 Nonanedioate 9.4 8.gC 9.7 AlO,l3 3 11.tY 5.7 Decanedioate 13.2 A11.14 2 17.4 3.8 Undecanedioate 17.7 Malonate 1.4 Samples (2-3 pl.) were injected into a 5-ft. packed helical column cona An “Aerograph” unit with a catherometer detector was used. Methyl taining a polyester stationary phase (“LAC-446”). Helium was used a t a constant flow rate (rotameter reading, 85.0 mm.). Column temperature: 195’. ester of Hormel “high purity” linoleic acid. A8.11
Heptanoate Hexanoate Pentanoate Butanoate
4.28 2.60 1.6 0.91
4.32 2.55 1.6 0.92
due to deterioration on etorage, or t o partial isomerization of the a c i d P during the 2-hr. period on the column is not known. Cleavage Analysis of Skipped Unsaturated Acids 1X.-The procedure given below for the ozonolysis and analysis of methyl cislO,cis-13-0ctadecadienoatewas followed for all the methyl esters. Ozonized oxygen was passed through a solution of 0.2 g. of the methyl ester of 10,13-octadecadienoic acid (IX, m = 3, n = 8 ) in 10 ml. of pure chloroform in a bath a t -18”. The gas flow was interrupted when the emergent gases developed a brown color in an aqueous potassium iodide solution, and the reaction mixture was then allowed to stand a t room temperature for 0.5 hr. Solvent was removed a t 30” by distillation in vacuo, and the residual oil was treated with a slurry of freshly prepared silver oxide (0.9 g.), 10 ml. of water, and 1.6 ml. of 10% aqueous sodium hydroxide in several portions. The mixture was stirred vigorously and heated a t 9@95” during and for 1 hr. after the addition.” Hydrochloric acid (20%) was added to pH 2. The mixture was extracted several times with ether, and the combined extracts, after one rinsing with water, were dried with magnesium sulfate. Treatment of the solution with diazomethane, as described above for the preparation of the methyl octadecadienoates, esterified all (35) The methyl esters of polyunsaturated acids suffer no significant change during gas-liquid chromatography at 197’ with Apiezon hl as the stationary phase [W. Stoffel. W. Insul, Jr., and E. H. Ahrens, Jr., Chem. Abstr., 63, 3973 (1959); Proc. SOC.Ezpll. B i d . Med., 99, 238 (lQ58)I.
free carboxylic acid groups. When volatile material was removed a t steam temperatures under a moderately reduced pressure, 0.2 g. of residue consisting largely of methyl valerate and dimethyl sebacate remained. Yields ranging from 80-957, were obtained a t this point. No dimethyl malonate was recovered. Table X summarizes the results of gas-liquid chromatographic analysis of the mixture. One or more minor peaks revealed the presence of lower, homologous diesters. The diester tracings were used to obtain the percentage impurities listed in Table X . Whether the values obtained in this way in fact reflect the degree of inhomogeneity in the 18-carbon acids, or whether the values are artifacts or the consequence of the degradation of initially homogeneous cleavage productsz5was not determined.
Acknowledgment.-We wish to thank Abbott Laboratories, North Chicago, Ill., for a grant which supported this work in part. We also appreciate the cooperation of Drs. S. R. Lipsky, R. A. Landowne, and R. K. Beerthuis in carrying out gas-liquid chromatographic analyses of our final products, and of Mr. James Pavlin in coaching us on the handling of isobutylaluminum compounds. Hercules Powder Company very kindly provided a generous sample of triisobutylaluminum.
Yerba Buena. 11. The Identification of Micromerol as Ursolic Acid’ GEORGEH. STOUTAND KENNETH L. STEVEXS Chemistry Department, University of Washington,Seattle 5, V’ashington Received August 31, 1962 It has been shown that micromerol is mainly ursolic acid, mixed, in some cases, with smaller amounts of oleanolic acid. The three-dimensional crystal structure of methylmicromerol bromoacetate (methyl ursolate bromoacetate) has been determined in the course of this study, providing an independent confirmation of the stereochemistry of the a-amyrin system.
Among the numerous products isolated by Power and Salway2from yerba buena (Satureia douglassii) were two colorless alcohols to which were assigned the formulas C33H6203e2Hz0 and C30H4604‘2HzO and the names micromerol and micromeritol. Of these, micromerol was present in significantly greater quantities. Although the proposed formulas were not in agreement, the properties described for these compounds suggested strongly that they were triterpene carboxylic acids. I n the course of our isolation of xanthomicrol’ we obtained by cooling the crude, concentrated ethereal extract of yerba buena a copious greenish white precipitate. This showed on thin - layer chromatography (t.1.c.) two major components having the color reac(1) Previous paper, G. H. Stout and V. F. Stout, Tetrahedron, 14, 296 (1961).
(2) F. B. Power and A. H. Sslway, J. A m . Chem. Soc., 30, 251 (1908).
tions of triterpenes.* The less polar material was present in considerably larger amounts and proved to be a colorless hydroxy acid, whose properties were in agreement with those reported for micromerol. Our sample and its derivatives, however, gave analyses compatible with the formula C30H4*03,Le., a mono hydroxylated triterpene acid. Methylation of micromerol with diazomethane gave a methyl ester which showed the same melting points for the hydrated and dried forms as reported by Powers. Since the ester appeared to be rather more easily crystallized and purified than micromerol itself, most of our material was isolated in this form by chromatography (3) We wish t o thank Mr. Ench Gsuglitz, U. S. Bureau of Commercial Fisheries, Seattle, for first pointing out t o us the striking advantages of this technique. (4) C. R. Noller, R. A. Smith, G . H. Harris, and J, W, Walker. J . Am. Chem. Soc., 64, 3047 (1942).
